We performed a randomized double-blind trial to determine the usefulness of early methylprednisolone therapy for patients with pulmonary failure. We selected 81 acutely ill, mechanically ventilated patients at high risk for adult respiratory distress syndrome (ARDS). Thirty-nine patients received methylprednisolone, 30 mg/kg, every six hours for 48 hours; 42 patients received mannitol placebo. All patients were given a positive end-expiratory pressure of 5 cm H2O, monitored with pulmonary artery catheters, and treated for their primary disease processes. Twenty-five steroid-treated patients (64%) and 14 placebo-treated patients (33%) developed ARDS. Early infectious complications occurred in 30 steroid-treated patients (77%) and 18 placebo-treated patients (43%). There were no significant differences in factors predisposing to ARDS, ventilatory requirements, or days of intensive care. These results do not support the use of methylprednisolone for ARDS. Steroids failed to improve pulmonary function and were associated with an increased infection rate. Intensive pulmonary and general supportive care remain the preferred therapy for ARDS.